Informativeness of bank financing announcements and relationship banking: Models and tests

This Ph.D. thesis deals with issues in the banking area via four chapters and two essays involving two original theoretical models and a set of empirical tests. The first chapter is an introduction to the issues addressed and methodologies used in this thesis. The next two chapters consist of two self-contained essays where our theoretical models and empirical results are presented. The last chapter provides appropriate conclusions.; Essay One is entitled "Do Banks Always Make Informed Loans?---Model and Some Tests", and Essay Two "Can Relationship Banking Be Undesirable to Firms?" While the problems examined in these two essays differ, both concern information asymmetry and involve game-theoretic analyses. In the last sections of Essay One, we test various implications of the model developed in the same essay. The following is a brief description of the two essays.; In Essay One, we model a signaling game between banks and financial analysts who follow their stocks, and characterize a perfect Bayesian equilibrium with a view to examine how diversely able (heterogeneous) banks lending to the same customer may impact the informativeness of their loan decisions. Distinguishing between H-banks and L-banks, those with high or low lending ability, respectively, we illustrate how two factors can work to bring an L-bank to act rationally against its valuable signal about its borrower in its loan decision-making (called uninformed lending), thereby blunting the information externalized to the public. The factors are: (1) the asymmetry arising from banks having private information about their own lending abilities; and (2) the concern of the banks to have financial analysts assess favorably such abilities. We identify the equilibrium conditions under which L-banks would practice uninformed lending. The analysis yields some implications that are tested empirically using a sample of Canadian bank loan announcements and withstands rejection. More interestingly, the results shed some light on various puzzles documented in the literature. In particular, while we confirm the well documented evidence that bank financing conveys information about the borrowers, we also find support for the key predictions that (1) the bank-financing announcements cause market reactions that are the strongest for deals involving single banks only, as opposed to multiple banks; (2) market reactions to bank financing are inversely related to the borrower firms' prior prospects, indicating that bank financing agreements convey much more information when uninspiring firms borrow. Our main conclusion from this study is that while banks generally perform informed lending and externalize information about borrower firms, there are circumstances under which the market does not react as much to banks' lending announcements, suggesting that the market seems to recognize the incidence of what we characterize as "uninformed lending" by banks and discount it accordingly.; In Essay Two, we model a lending game between two banks that differ not only in proprietary information about a borrower firm's creditworthiness, as is commonly the case in the existing literature, but in monitoring capability to improve the firm's creditworthiness as well. Within this setting, we demonstrate that when a bank has informational advantage over its rivals, the rival banks, even with superior monitoring capability, may not be able to get any profitable lending business. The borrower firm is thus deprived of a potentially more beneficial loan. After solving for the model's equilibrium, we draw implications of our analysis and discuss their relevance in practice. Given the consistency between implications of this analysis and the rationale behind two major regulatory policy changes in recent history within the banking industries in both Canada and the U.S., we estimate our model proposes a fairly convincing argument.; Overall, we estimate that our novel theoretical results along with the empiric...